Purpose To determine whether quantification of liver surface nodularity (LSN) provides an estimate of the presence of clinically significant portal hypertension (CSPH) in patients with cirrhosis. Materials and Methods This retrospective study included a training cohort (n = 189) and separate external validation cohort (n = 78), both composed of patients with cirrhosis who underwent abdominal CT and hepatic venous pressure gradient (HVPG) measurement between 2010 and 2016. The LSN score, liver and spleen volumes, liver-to-spleen volume ratio, platelet count to spleen diameter ratio, Iranmanesh score, aspartate amino transferase-to-platelet ratio index, and Fibrosis-4 index were derived from CT images and serum laboratories. The accuracy of the various tests for predicting CSPH was evaluated with area under the receiver operating characteristic curve (AUROC) and compared by using the DeLong test. Student t test and Pearson correlation coefficient were used. Results One hundred eighty-nine patients were analyzed (119 men [mean age ± standard deviation, 57 years ± 11; range, 29-81 years] and 70 women [mean age, 61 years ± 10; range, 34-83 years]; overall mean age, 58 years ± 10; range, 29-83 years). A total of 102 patients (54%) had CSPH. LSN score correlated with HVPG (r = 0.75; P < .001). Patients with CSPH had a higher LSN score than did those without CSPH (3.2 ± 0.6 vs 2.4 ± 0.3; P < .001). A cutoff value of 2.8 had a positive predictive value of 88% for CSPH; the AUROC of LSN was 0.88 ± 0.03. This was higher than that of other available noninvasive tests (DeLong, all P < .001). In the validation cohort, LSN score of 2.8 had a positive predictive value of 86% for CSPH; the AUROC was 0.87 ± 0.04. Conclusion The CT-based liver surface nodularity score demonstrated high diagnostic performance for detecting clinically significant portal hypertension and outperformed multiple other noninvasive tests. © RSNA, 2018 Online supplemental material is available for this article.